Detecting near-duplicate document images using interest point matching | IEEE Conference Publication | IEEE Xplore